Live Code Update for IoT Devices in Energy Harvesting Environments

The number of Internet of Things (IoT) devices is exhibiting explosive growth. These devices are often closely coupled to the physical world, and may harvest energy as a power source, which imposes particularly stringent operating constraints. Like any programmable system, IoT devices may need software updates to fix bugs, add functionality, or improve computational capability. This paper proposes novel strategies to update deployed code for IoT energy-harvesting devices based on in-place code updating and code trampolines, which effectively eliminate system down time and minimize resource demands for updates. We show that our schemes, on average, reduce the number of nonvolatile memory writes by 99% and code transmission cost by 78%.